In May, Ancestry changed the chip that they use for autosomal DNA processing and comparison. They removed roughly 300,000 of their roughly 682,000 locations and replaced them with medical SNPs. That means that people who tested before the middle of May, 2016 are only being compared to a little more than half of the SNPs on the chip of the people who tested on the V2 chip after the middle of May, 2016.
Clearly there are going to be some differences in matches reported. Ancestry said they should be minimal, but I must have some Missouri blood someplace, because I wanted to see for myself. I ordered a V2 test to see just how the V1 and the V2 tests compare.
I am specifically interested in ethnicity percentages and match numbers. But first, let’s step through the order process.
Ordering at Ancestry
Ordering a second kit was amazingly simple – done just by clicking on my current account “Order a new kit.” They keep my credit card information on file, so literally it was a one or two click process. Unfortunately, what they didn’t do was to have me read all of the Terms and Conditions and small print when I ordered, so by the time the kit arrived, and I was already financially invested, there was little I could do about the Ts&Cs if I didn’t like them. I strongly suspect most people don’t read the fine print, because at that point, it doesn’t matter since they’ve already paid for the kit and made the purchase decision. And let’s face it, you’re excited about the kit arriving and want to take the test.
After my kit arrived, I had to activate the test, and of course, I got to do some clicking and answer some questions. Let’s walk through that process, because it has changed since I ordered my original kit several years ago.
When you click the box that says “I have read the Terms and Conditions,” actually read the Terms and Conditions. It’s unfortunate that you don’t see the Terms and Conditions until AFTER you’re purchased this product – because the contents of the Terms and Conditions might well affect your decision about whether to purchase this DNA test or not. Maybe that’s why it’s here and doesn’t appear during the purchase process!
Here’s a link to the Terms and Conditions.
Please take note specifically of the following paragraph from the Terms and Conditions document:
By submitting DNA to AncestryDNA, you grant AncestryDNA and the Ancestry Group Companies a perpetual, royalty-free, world-wide, transferable license to use your DNA, and any DNA you submit for any person from whom you obtained legal authorization as described in this Agreement, and to use, host, sublicense and distribute the resulting analysis to the extent and in the form or context we deem appropriate on or through any media or medium and with any technology or devices now known or hereafter developed or discovered. You hereby release AncestryDNA from any and all claims, liens, demands, actions or suits in connection with the DNA sample, the test or results thereof, including, without limitation, errors, omissions, claims for defamation, invasion of privacy, right of publicity, emotional distress or economic loss. This license continues even if you stop using the Website or the Service.
Note that the Terms and Conditions then links to the Ancestry Privacy Statement, which by implication is part of the Terms and Conditions, so read that too. And that statement is different from the AncestryDNA Privacy Statement, so you’ll want to read that as well.
Please note specifically in the Ancestry DNA privacy statement, the following paragraph, 6-i:
Non-Personal Information also includes personal information that has been aggregated in a manner such that the end-product does not personally identify you…
Because Non-Personal Information does not personally identify you, we may use Non-Personal Information for any purpose, including sharing that information with the Ancestry Group Companies and with other third parties. In some instances, we may combine Non-Personal Information with personal information (such as combining your name with your geographical location). If we do combine any Non-Personal Information with personal information, the combined information will be treated by us as personal information, as long as it is combined, and its use by us will be subject to this Privacy Statement.
Ancestry then asks you about Research Project Participation, which is a specific authorization for third party research projects that is different from the above.
You can read the entire Informed Consent document here.
How many people do you think actually read, and understand, all 4 documents hot linked above? If you do, that makes 2 or 3 people that I know of. If you have insomnia, these documents will cure it, guaranteed:)
I’m glad to see Ancestry encouraging people to link to trees. Now if testers would just make those trees public instead of private.
I did link to my tree during the activation process, but when my results came back, my tree was not linked. Be sure to check…otherwise you won’t have any Circles or Shared Ancestor Hint leaf matches which means your DNA matches and you share a common ancestor in your trees.
A couple of steps didn’t work correctly, but I was still able to register the kit.
Another item, which I think is important and I don’t believe was reflected in the Terms and Conditions verbiage is that Ancestry kits are now being processed by an outside lab, Quest Diagnostics.
A few weeks later, my V2 results were returned, so let’s take a look at how they compare to V1.
V1 versus V2 Match Results
It took one day short of a month, after my test reached Ancestry, for my results to be returned.
Given the rather dramatic change in the number of genealogy SNPs on the Ancestry chip between V1 and V2, and given that only about half of the locations are the same between the V1 and V2 chips, I expected significantly fewer matches on the V2 chip than on the V1 chip. In other words, I didn’t expect that the V2 chip would be nearly as effective in matching the V1 test takers, because those two chips only shared about half of their locations.
There were more V1 matches, but not nearly as many as I expected.
*Leaf matches are Shared Ancestor Hints that mean you match someone’s DNA who also has a common ancestor listed in their tree. This is by far the most useful DNA tool at Ancestry.
I need to confess here that the matches I’m actually the most interested in are those “Shared Ancestry Hints” matches, with a leaf, because the common ancestor is identified for the matching pair of people, unless one of the two has a private tree. Then the non-private tree user cannot see the private tree’s ancestors, which means that you cannot determine the common ancestor with someone who has a private tree.
The 19 Circles are the same Circles for both kits, which is what I expected. This also tells me that the missing matches weren’t that critical match that made the difference between a Circle being formed, and not.
Unfortunately, there is no good way to print or download your list of matches, at least that I have been able to discover. I used (ctrl+P) to print all 7 pages of leaf hint match, which was 12 printed pages for each match page, in case you’re so inclined. I then compared the V1 to the V2 matches manually. Yes, this was a huge pain, spread across 84 pages. However, I really wanted to see if the V1 kit leaf matches were the same as the V2 leaf matches. These should be a good representative sample of the rest of the matches, and I’m not about to manually compare 15,000 matches.
Of the old V1 kit matches, 7 matches were present on the V1 list and absent on the V2 list, including my last two “lowest confidence” matches who were obviously teetering on the threshold – and some of those missing SNPs were just enough to push us below the threshold, so we are not considered a match on the V2 chip.
For the new V2 kit, only one match was present on the V2 list and absent on the V1 list. Apparently that one kit’s critical matches were in the area of the medical SNPs and the genealogical SNPs alone (if there were matches in that area) did not cause the kit to rise above Ancestry’s matching threshold. Unfortunately, without a chromosome browser, we can’t see anything about the locations of the matches on our chromosomes.
About 2.5% of the matches were absent in the V2 test when compared to the V1 test. However, the net difference of 5 was not reflective of the matches being the same. A total of 8 were absent from the other test, in total.
Other than these 8 kits, the rest were the same matches in both kits. I would suspect that the matching percentage of about 97.5% would hold for the total matches as well.
Of the highest confidence matches, all of the matches were present. The match order was often significantly different, indicating that the reduced SNP count did matter in terms of how well they matched, but did not reduce the match enough to cause them to drop off the match list – except for those 7 of course. As expected, the V1 kit did out-perform the V2 kit, but not by a lot.
NADs are New Ancestor Discoveries, which are inappropriately named.
It’s interesting that the new V2 kit has no New Ancestor Discoveries. I checked several times over two or three weeks, thinking that some might appear. That’s actually fine with me, because, as I’ve written before, NADs have proven to be entirely useless. Still, if I were a V2 test taker, especially an adoptee or someone with unknown parentage, I would want every hint I could get. In the past few days/weeks, the same NADs on the V1 account have been coming…and going…and coming…and going. If you don’t have NADs, and you want NADs, give Ancestry’s customer support a call and ask them to kick the tires for you. Lack of NADs could be a bug.
**The day I did the initial comparison between the V1 and V2 kits, I had 3 NADs on the older V1 kit. Two days after I did the initial comparison on the V1 kit, I had 7 NADs (which remain 3 weeks later) and still zero on the newer V2 kit. Today, the NAD total is 8 on the old V1 kit and still zero on the new V2 kit.
V1 versus V2 Ethnicity
The V1 versus V2 Ethnicity is really nothing to write home about. There was a very slight difference between two categories, by 1% each. Scandinavia, where I have no documented lines, moved from 10 to 11% and Great Britain, where I have multiple lines, moved from 4% to 3%. Go figure.
It’s somehow ironic that my trace regions include 3% in Great Britain and 2% in Ireland, where I have multiple documented lines, and the same amount, 2% in Italy and Greece combined where I have absolutely no connection at all.
As I’ve said before, about all of the testing companies, these ethnicity tests tend to be relatively reliable between continents, meaning Europe, Asia, Africa and Native American – and much less reliable within continents. Don’t be trading in your kilt (or anything else) based on these kinds of tests.
I was really quite pleasantly surprised that the matching difference wasn’t greater between chips. And truthfully, the matches I’m the most interested in are my closest matches, because they are the matches with whom I’m most likely to be able to identify a common ancestor – and my Shared Ancestor Hints leaf matches, because a common ancestor is already identified. All of my close matches were present in both kits – probably because losing some matching segments didn’t affect the fact that we do match. Most of my Shared Ancestor Hints were retained too. The matches that were lost tended to be the lower matches, based on Ancestry’s highest to lowest matching order.
Losing just under 2% of more than 15,000 matches isn’t anything I’m going to lose any sleep over. Losing 2.5% of my leaf matches isn’t anything I’m going to lose sleep over either, although those certainly do hold more promise than non-leaf matches. I would like those additional 8 leaf matches not present in the other kit, but again, I wouldn’t lose sleep over those either.
The net-net of this is that if you have already taken the V1 test, before May of 2016, you don’t need to order the V2 test. The V2 test is slightly less productive, but all in all, it’s still of the same approximate quality as the V1 test – except for those NADs.
If I had already tested on the V1 kit, I certainly would not pay an additional $99 for 1 additional Shared Ancestor Hint leaf match that I’d have to manually compare with the other kit to find – and I would have to maintain that duplicate comparison into the future. I went through that process for this article, but had I been doing this just for myself and known the outcome in advance, I truthfully wouldn’t have bothered. It’s a lot of work for very little return.
The differences in terms of matches, ethnicity and circles are minimal, and the V2 test received slightly fewer matches in total, slightly fewer leaf matches and no NADs – so there would be absolutely no benefit in retesting on V2 if you’ve already tested on V1 – aside from 1 match that you’ll have to manually compare to find. I’m glad I took the original V1 test, because it does fare somewhat better overall, but not enough to make a lot of difference.
I have been pretty unhappy with some of Ancestry’s past choices and changes, to put it mildly, but this time, Ancestry seems to have done this right. I wish Ancestry hadn’t changed chips at all, because their motivations are entirely self-serving and the chip change doesn’t benefit the genealogist at all. However, in terms of how Ancestry handled this chip conversion, and compared to 23andMe’s disaster, Ancestry hit a home run. The change may not benefit Ancestry’s customers, but it also doesn’t damage them (much) or impact their ability to utilize the testing and matching for genealogy – which is why they purchased the test in the first place.
Ancestry was correct when they said that the V2 chip wouldn’t affect matching much with the V1 chip customers, and that there was no need for V1 customers to purchase a new V2 test.
Now, if Ancestry would just implement a chromosome browser so we can see how and where we match people – we would all be really happy campers!!! Yes, I know, Hades has not yet frozen over…but hey…winter’s coming and hope springs eternal.